This study develops a research protocol and new instrument for measuring the characteristics and content of group therapy in alcohol treatment programs. As health services researchers attempt to apply rigorous scientific methods to measure the effectiveness of alcohol treatment, it becomes essential to reliably and comprehensively characterize the ingredients of treatment. While instruments to measure program characteristics and particular domains of group treatment have been developed, no instrument exists that measures across the domains of group treatment, the major modality in alcohol treatment programs today. The project develops a structured instrument and training manual, and evaluates the capabilities of instrument self-administration by clients, therapists, and outside observers of alcohol therapy groups. Instrument development pays special attention to developing items that are generalizable to group treatment experiences relevant to both men and women's needs. Phase I involves a literature/instrument review and expert review panel to develop an initial item bank covering 5 domains of group alcohol treatment included in our conceptual model. We also conduct a series of structured, in-depth interviews and protocol analyses with male and female clients in alcohol treatment and staff members to further specify and abbreviate items, as well as to improve instrument coverage. Phase 2 of the research focuses on field-testing the instrument's reliability and validity, examines the effects of using alternative methods of instrument administration, and conducts analyses to further refine the instrument's structure with the goal of achieving the minimum set of items that provide full coverage of the group treatment experience. We sample group sessions from across a range of public sector, HMO and private outpatient alcohol treatment programs. To capture multiple perspectives on treatment, to establish inter-rater reliability, and to compare alternative methods of instrument administration, Phase 2 will also include data collection from therapy group leaders and independent group observers. The result is intended to be a brief, usable method for measuring the topical content, level of participation, treatment philosophy, group features, and group leader characteristics of group treatment. The instrument will facilitate not only our planned research focusing on health service delivery related to alcohol treatment, but also future studies of treatment efficacy, treatment matching, and program development, evaluation, and management in the clinical practice of alcohol treatment.